CODEWITHSUNDEEP
pythontensorflow
John
John

Reputation: 93

Subtracting matrix elements

Suppose I have one 4x3 matrix. I want to subtract every element in that matrix from each other.

I've looked around TensorFlow documentation extensively (and on SO) and noticed that there is a tf.subtract operator. In adding, I know that there is tf.add_n operator which adds all input tensors. I'm new to TensorFlow and was wondering: is there such a subtraction operator which subtracts all input tensors and if not, can you please provide an example of the fastest way to do so?

Example matrix tensor: [[0.10, 0.20], [0.20, 0.40]] so expanded as:

0.10 0.20
0.20 0.40

Desired subtraction: 0.10 - 0.20 - 0.20 - 0.40 with desired output as: -.7

Upvotes: 0

Views: 261

Answers (1)

John R
John R

Reputation: 1508

What you’re describing is the opposite of the sum, substituting the opposite of the first element of the first vector.

tf[0][0] = tf[0][0] * -1

And then use 

tf.reduce_sum() * -1

Upvotes: 1

Related Questions